@q >each follow gets their own room
>my room always at the top of your list after i retweet 100 fox girl posts
LMAO
>my room always at the top of your list after i retweet 100 fox girl posts
LMAO
Actually true
@q PING another fox girl post
how come your reply to your post mentioning me didnt notify me?
CLAUDE FIX IT
how come your reply to your post mentioning me didnt notify me?
CLAUDE FIX IT
I replied to my message, which I think was a bad move. I should have just kept it at the thread root.
@q wait does this thing map to matrix threads? 😬
@q if that was an edit to your previous reply, there is an edit function in pleroma
not sure how it works but you can somehow try to map it so it works on both matrix/fedi
not sure how it works but you can somehow try to map it so it works on both matrix/fedi
That was a redact and resend. I think I do have edits working. I'll try adding something.
EDIT: nigger!!!
bronze Almost had media with captions working and hit my limit again 😭
@q i hate how it locks you out even if you're a pay cuck
like whats even the point? i should be able to abuse claude as much as i wish if im a paying customer
someday ill finish the MMD/vrchat model to vtuber program
like whats even the point? i should be able to abuse claude as much as i wish if im a paying customer
someday ill finish the MMD/vrchat model to vtuber program
@bronze @q so, I subscribed to Claude. Saw that Fable has a few days for using as a subscriber, and they also lock you out after a few messages each 5 or 6 hours.
So the first thing I did was say "hey, here are my current computer specs and operating system. Help me build the best LLM I can, running completely in my own hardware, leveraging my GPU, CPU and RAM, and it has to fit these requirements and basically replace you.". Then I put it on Fable and cranked it to max settings so it would use a bazillion tokens thinking about it.
It explained all the options and how they work, then created a document outlining an 8 step plan to do it, each one building on the previous one, and each one allowing for independent testing to be assured all the parts are moving correctly.
I'm gonna mess with it this weekend. It seems it won't be absolutely perfect, but it's literally the best I can do and will leverage my whole computer by using MoE models (I have a beefy CPU and 128GB of RAM, but my GPU is just a RX 9070XT with 16GB VRAM, so I can't just throw a 40B model in it an call it a day). And, of course, it will run locally so I won't need to be cucked by these companies again.
So the first thing I did was say "hey, here are my current computer specs and operating system. Help me build the best LLM I can, running completely in my own hardware, leveraging my GPU, CPU and RAM, and it has to fit these requirements and basically replace you.". Then I put it on Fable and cranked it to max settings so it would use a bazillion tokens thinking about it.
It explained all the options and how they work, then created a document outlining an 8 step plan to do it, each one building on the previous one, and each one allowing for independent testing to be assured all the parts are moving correctly.
I'm gonna mess with it this weekend. It seems it won't be absolutely perfect, but it's literally the best I can do and will leverage my whole computer by using MoE models (I have a beefy CPU and 128GB of RAM, but my GPU is just a RX 9070XT with 16GB VRAM, so I can't just throw a 40B model in it an call it a day). And, of course, it will run locally so I won't need to be cucked by these companies again.
@VooDooMedic @q @bronze yeah, that's the whole idea.
